Clean up Cliquer's spkg-install

[sagetrac-drkirkby]

The spkg-install file for Cliquer could do with a lot of cleaning up. Some examples of the problems are:

• There are flags set for the C++ and Fortran compilers, though the code is only C.
• There's code to check for a mix of Sun and GNU compilers, which is pointless since there is only 1 compiler used.
• Lots of unnecessary environment variables are set.
• Use make instead of $MAKE.
• Runs tests in spkg-install instead of spkg-check.

None of these issues are currently causing any problems, but should be resolved at some point

Depends on #14349

CC: @nathanncohen @sagetrac-mvngu

Component: packages: standard

Author: Jeroen Demeyer

Branch: 2ffeeb7

Reviewer: Jean-Pierre Flori

Issue created by migration from https://trac.sagemath.org/ticket/9870

[nexttime]

comment:1

To tranquilize Dave :-)

I've made several comments at tickets related to Cliquer, e.g. #9833 and #9871.

I think this should then be a follow-up of #9871, despite the numbers.

[nathanncohen]

comment:3

Hello !

There is one technical question you can however answer. Do we need the binary file "cl" so it can be executed from the command line, or is the library libcliquer.so sufficient?

No, the cl file is not used, and this is precisely what the Sage code included in cliquer is useful for : directly calling the library's functions with a Graph object using the Graph structure it expects to find, without having to create an ugly temporary file to call the executable on it :-)

Nathann

[sagetrac-drkirkby]

comment:5

Leif, you were keen to take ownership of this. Has anything happened with it?

Dave

[sagetrac-drkirkby]

comment:6

I ask once again. Any chance of you sorting this out Leif, since you wanted to take ownership of it.

Dave

[jdemeyer]

comment:7

See #11227 for another issue with cliquer, unrelated to this ticket.

[jdemeyer]

Author: Jeroen Demeyer

[jdemeyer]

comment:9

I'll tackle this.

[jdemeyer]

Branch: u/jdemeyer/ticket/9870

[jdemeyer]

Commit: 5f55f5b

[jdemeyer]

New commits:

f55a304	python-sage work without cliquer patching (#14349)
aecdf29	module_list: add cl.c to cliquer.so sources (#14349)
25edb2f	trac #14349: rebase on 6.1.beta4
6253068	trac #14349: move cl.* to cliquer/
9f04e98	Merge remote-tracking branch 'origin/develop' into HEAD
5f55f5b	Clean up cliquer package

[jdemeyer]

Dependencies: #14349

[jpflori]

Changed commit from 5f55f5b to 2ffeeb7

[jpflori]

Reviewer: Jean-Pierre Flori

[jpflori]

comment:14

Don't know why trac was not able to merge the branch as I encountered no conflicts when merging it.

I've made a few changes to make the script more like the "new" scripts as described in the dev guide, you may not be happy with them and are very welcome to change them back.
Otherwise let's positively review this ticket.

Anyway, I'm going to autotoolify cliquer, post it in a follow-up ticket and suggest it upstream.
The Makefile is just too awful right now.

---

New commits:

92dd8c8	Merge remote-tracking branch 'trac/develop' into ticket/9870
2ffeeb7	Slight changes to cliquer install script.

[jpflori]

Changed branch from u/jdemeyer/ticket/9870 to u/jpflori/ticket/9870

[jpflori]

comment:15

I've sent an email upstream about the possibility of releasing an official version including some of the Debian's patches (http://anonscm.debian.org/gitweb/?p=debian-science/packages/cliquer.git;a=tree;f=debian/patches;hb=HEAD) and an autotoolified build system.

[jdemeyer]

comment:16

I think set -e is a good thing and should be used more. So I reverted that change.

[jdemeyer]

Changed branch from u/jpflori/ticket/9870 to u/jdemeyer/ticket/9870

[vbraun]

Changed branch from u/jdemeyer/ticket/9870 to 2ffeeb7

[jpflori]

Changed commit from 2ffeeb7 to none

[jpflori]

comment:19

FYI, upstream has nicely answered my questions.
They don't really mind cliquer being distributed in different forms in various places and don't plan on releasing any new upstream version "in the next ten years or so (or ever?)".